摘要
为了快速测定樟树(Cinnamomum camphora)的含油率,建立了樟树含油率近红外光谱测定模型。利用台式近红外光谱仪测定200份樟树叶样品的光谱信息,水蒸气蒸馏法提取精油并计算含油率,然后结合化学计量学方法进行构建,最后使用20份未知样品对该模型进行外部验证。将异常样品剔除后,最终选取196份样品参与模型构建。对不同建模方法、建模波段以及预处理所构建模型的效果进行比较,从而选定最佳建模波段为1000.00~2000.00 nm,Segment 5 Gap5的3点二阶泰勒求导作为最佳预处理方式,并结合主成分回归算法构建出最佳模型。最佳模型的校正集相关系数(R_(c))为0.9561,均方根误差(R MSEC)为0.4079;交互验证集相关系数(R_(v))为0.9479,均方根误差(R_(MSEV))为0.3699。外部验证结果表明:未知样品的化学测定值与模型预测值之间不存在显著差异。所构建的模型具有优良的预测性能,能够应用于大批量樟树含油率的快速测定。
In order to rapidly determine the oil content of Cinnamomum camphora,a near-infrared spectroscopy determination model for the oil content of C.camphora was established.The spectral information of 200 camphor leaf samples was measured using a benchtop near-infrared spectrometer.Essential oils were extracted by steam distillation,and the oil content was calculated.Then,the model was constructed in combination with chemometric methods.Finally,20 unknown samples were used to externally validate the model.After removing the abnormal samples,196 samples were ultimately selected to participate in the model construction.The effects of models constructed by different modeling methods,modeling bands,and pretreatment methods were compared.As a result,the optimal modeling band was determined to be 1000.00-2000.00 nm,and the 3-point second-order Taylor derivative with Segment 5 Gap5 was selected as the best pretreatment method.The best model was constructed by combining with the principal component regression algorithm.The correlation coefficient(R_(c))of the calibration set of the optimal model was 0.9561,and the root mean square error(R MSEC)was 0.4079.The correlation coefficient(R_(v))of the cross-validation set was 0.9479,and the root mean square error(R_(MSEV))was 0.3699.The results of external validation showed that there was no significant difference between the chemical determination values of the unknown samples and the predicted values of the model.The constructed model has excellent predictive performance and can be applied to the rapid determination of the oil content of a large number of C.camphora.
作者
涂白连
江斌
伍艳芳
郑永杰
刘新亮
张月婷
徐海宁
曾常金
Tu Bailian;Jiang Bin;Wu Yanfang;Zheng Yongjie;Liu Xinliang;Zhang Yueting;Xu Haining;Zeng Changjin(Jiangxi Academy of Forestry,Nanchang 330032,P.R.China;Forestry Farm of Yongfeng County)
出处
《东北林业大学学报》
北大核心
2025年第7期118-123,共6页
Journal of Northeast Forestry University
基金
江西省林业科技创新专项项目(创新专项[2021]15号)
江西省林业科学院基础研究与人才科研专项项目(2023522703)
江西省重点研发计划项目(20203BBF62W010)。
关键词
樟树
含油率
近红外光谱
Cinnamomum camphora
Oil content
Near-infrared spectroscopy
